There is an unmet need to obtain natural product-derived medicines in a scalable and reliable manner. Natural products are small molecules produced by biological systems, but often in low quantities. Metabolic engineering promises to move the biosynthetic pathways of these medicinal small molecules from their native producers into heterologous hosts such as bacteria and yeast, which can be cultured in large-scale, low-cost, industrial processes. The long-term goal of this proposal is to address the grand challenge in metabolic engineering, namely the extraordinary complexity of metabolism in the cells, by insulating virtually any desired pharmaceutical biosynthetic pathway from the rest of cell metabolism, so that the former can be studied and optimized effectively. The central hypothesis is that this goal can be achieved by establishing an unnatural redox cofactor system (uRedox) in vivo to power the desired biosynthetic pathways. This design is inspired by Nature: Catabolism and anabolism, two opposing metabolic systems responsible for breaking down and building up cell components, respectively, are insulated from each other because they each have a designated redox cofactor, NAD and NADP, respectively. The scientific premise of this hypothesis has been demonstrated: when using nicotinamide mononucleotide (NMN) as an unnatural redox cofactor in vivo, only the engineered, productive reaction for a pharmaceutical was active, while all other interfering reactions in the host's complex metabolic network remained silent. To develop this prototype of uRedox into a truly universal technology, the following specific aims are proposed: (1) Develop a facile, high-throughput selection platform for obtaining NMN-dependent enzymes on demand and en masse. Specifically, Escherichia coli will be engineered so that only cells harboring active NMN-dependent enzymes can survive. In combination with the computational protein design pipeline that we established, this growth-based selection platform will allow rapid customization of uRedox to produce different pharmaceuticals; (2) Develop E. coli and Saccharomyces cerevisiae, two most important industrial model hosts, with a built-in NMN pool. Some other organisms can accumulate NMN inside the cells, and their NMN biosynthetic pathways will be transplanted into E. coli and S. cerevisiae. The two engineered hosts will serve as chassis for implementing uRedox; (3) Apply uRedox to address a major need in metabolic engineering: preserving aldehydes. Many medicinal compounds or their biosynthetic intermediates are aldehydes, which cannot stably exist in cells because they are modified by the cells' numerous redox enzymes. As a proof-of-concept, uRedox will be used to render these aldehyde-modifying enzymes inactive all at once and thereby preserve a key aldehyde intermediate in the biosynthesis of benzylisoquinoline alkaloids (BIAs), a family of ~2,500 natural products including important antibacterial, antitussive, and analgesic drugs. The proposed approach is innovative because it directly targets life's universal metabolic infrastructure and therefore can have extremely broad impacts in biomedicine and synthetic biology.